The pulp and paper industry is the most energy intensive industry in Sweden. Most of the energy usedaims to supply the large amount of heat to the production processes that is required to produce pulp andpaper . By increasing the energy efficiency of paper mills, there is great potential to reduce productioncosts and reduce its environmental impact. This project involves mapping, analyzing and investigation ofmeasures that can contribute to increased thermal efficiency of an internal heating network at Holmen’spaper mill, Braviken. The heating network, called VVG, utilises excess heat from the pulp and paperproduction and supplies heat to production processes and premises at the factory. When the heat demandof the VVG-system exceeds the available recovered heat, heat from live steam is transferred. This projectaim at identifying measures with potential to reduce demand of live steam supply. The project started bymapping the entire VVG-system to identify which consumers are included and how it is connected. Thenthe consumers were studied and data was collected to determine their power and energy demands. Basedon the mapping and collected data, proposals for measures that have the potential to increase thermalenergy efficiency were noted which in turn can reduce the need for live steam supply. The measures havebeen focused on reducing the return temperature of the VVG-system and reducing the heat demand of thesystem. Reduced return temperatures are achieved by more efficient use of the existing heat flow. Thisleads to increased heat recovery potential which means that more excess heat can be utilised.The proposed measures that were identified as having potential to contribute to reduction of the returntemperatures of the VVG-system are as follow: • Utilising return flows with high flow rates and high temperatures to provide other consumers inVVG with heat. This allows the hot flow to be used instead of going directly into the return line. • Eliminate short circuits in the system where the supply line goes directly into the return line. Thesewere detected in the forms of three-way valves and flows through consumers that were not running. The proposed measures that have been identified as having the potential to contribute to reduction of theheating demand of the VVG are as follows: • Turning off flows passing through consumers that are not in use to reduce unnecessary temperaturedrops which will result in decreased power losses. • Preheating via residual flows at consumers where raw water tempered circuits is heated. As theraw water is cold during large parts of the year, a lot of heat can be saved by preheating it with aflow of lower energy quality than VVG. • Shut down the VVG flow through heat exchangers in the heat recovery system when the heatrecovery process not are running to avoid negative heat transfer. By implementing these measures, large improvements regarding the thermal efficiency of the VVG-system can be achieved at Holmen Paper Braviken. How much the return temperatures decrease and howmuch the heat recovery potential increases with the proposed measures has not been quantified during thisproject. However, the measures aimed at lowering the heat demand in the VVG-system can be quantifiedto contribute to an predicted energy saving of at least 2.8 GWh annually, but the investigations of themeasures indicates that significantly larger energy savings can be achieved. To ensure the potential of theproposed measures to contribute to the reduction of the demand for live steam supply, Holmen needs tocarry out further studies of the VVG system where this work is intended to form the basis for their furtherwork.

Objectif: La prévalence de la consommation de cannabis chez les Inuits du Nunavik est l’une des plus élevées au monde, particulièrement chez les adolescents, et les facteurs permettant d’expliquer ce phénomène demeurent mal connus. Selon Bronfenbrenner (1979), la famille est l’unité la plus proximale de l’adolescent et est susceptible d’influencer particulièrement ses comportements. L’objectif de cette étude est d’examiner de manière longitudinale l’association entre la consommation de cannabis de mères inuites et de leurs adolescents. Hypothèses : Une fréquence de consommation plus élevée chez les mères sera associée 1- à une initiation plus rapide au cannabis et 2- à une consommation plus fréquente chez leurs adolescents. Méthode : Les données utilisées proviennent de l’étude de cohorte prospective Nunavik Child Development Study (NCDS). Ce mémoire utilise les données de deux périodes de collecte, l’une survenue entre 2005 et 2010 (T1 : temps 1, jeunes 8-15 ans), et l’autre entre 2012 et 2016 (T2 : temps 2, jeunes 16-22 ans). Des entrevues structurées ont permis d’évaluer la consommation de cannabis des participants et d’autres variables concomitantes chez la mère à T1 et chez son adolescent à T2. L’échantillon analysé compte 184 dyades mère-adolescent. Des analyses de régression ordinale et de Cox ont été réalisées pour vérifier les hypothèses. Résultats : Aucune association significative n’a été identifiée entre la fréquence de consommation des mères et l’âge d’initiation au cannabis, de même que la fréquence de consommation de cannabis de leurs adolescents. Par ailleurs, un statut socioéconomique élevé et la fréquentation scolaire sont associés à un moindre risque de consommer fréquemment du cannabis chez les adolescents. Conclusion : L’association attendue entre la consommation maternelle et l’usage de cannabis de jeunes Inuits n’a pas été retrouvée. La recherche de prédicteurs à l’usage de substances reste d’actualité afin de minimiser le cumul de vulnérabilités que vivent les populations autochtones. / Aim: Adult and adolescent Inuit from Nunavik show some of the highest prevalence of cannabis use in the world, particularly for youth. Knowledge of the factors that explain this phenomenon in this region is lacking. According to Bronfenbrenner (1979), family is the closest unit for adolescents and is susceptible to particularly influence their behaviors. The aim of this longitudinal study is to examine longitudinally the association between maternal use of cannabis and the subsequent use of this substance by their adolescents in an Inuit population. Hypothesis: A higher cannabis frequency use among mothers will be associated 1- with more rapid initiation of cannabis, 2- and more frequent use among their adolescent. Method: This study was based on longitudinal data from the Nunavik Child Development Study (NCDS). This study used data from two follow-ups, between 2005 and 2010 (T1: time 1, youth: 8-15 years old), and between 2012 and 2016 (T2: time 2, youth: 16-22 years old). Structured interviews were used to assess maternal cannabis use and other concomitant variables at T1, and for adolescents at T2. Our sample included 184 mother-adolescent dyads. Hypotheses were tested using a Cox regression analysis and ordinal regressions. Results: No significant associations were found between mothers’ frequency of cannabis use and adolescents initiation age as well as regarding the adolescents’ frequency of cannabis use. Nevertheless, a higher socioeconomic status and going to school were associated with a lower risk of frequent cannabis use by the adolescents. Conclusion: The expected association between maternal cannabis consumption and cannabis use of young Inuit was not found. The search for predictors of substance use remains relevant to minimize the accumulation of vulnerabilities experienced by Aboriginal populations.

Master of Science / Department of Journalism and Mass Communication / Curtis Matthews / The objective of this study is to shed light on the interaction between the theory of selective exposure and social identity theory. Both of these theories describe aspects regarding how individuals expose themselves to information. However, the driving question of the study was to investigate the behavior of individuals whom experience cognitive dissonance from an inconsistent political message, but are receiving that message from a member of their perceived in-group. The expectation was that moderating variables, such as strength of identification with an ideology as well as the level of knowledge would impact who would engage in selective exposure and choose to disregard in-group cues, or choose inconsistent messages to stay true to in-group pressures. In a Qualtrics experiment, participants (n=189) were divided into different groups, attitude consistent and attitude inconsistent and were exposed to a series of tweets. Each tweet was politically charged, with the attitude-consistent group presenting consistent messages paired with in-group cues, and in the attitude-inconsistent group presenting a dissonant message was paired with the in-group cue, and visa versa. Two factors revealed themselves to impact results and message choice — knowledge and strength of identification. Results revealed that individuals who had a high level of knowledge chose the consistent message more often than those with low knowledge. The strength of ideology variable influenced differences in both the speed in which they made decisions on which tweet to select, as well as how quickly they identified with an ideology level. This result revealed that those who are strongly identified with an ideology make decisions regarding political messages and ideology faster than those who are weakly identified. The current study contributes to the plethora of literature regarding these two theories and the political science area of study by supporting knowledge as a moderating variable between cognitive dissonance and social identity pressures. It also provides insight into the trends and patterns that can arise when time/speed are utilized as a variable to shed light on group identification differences. The insights with the variables of time, strength of ideology, level of knowledge, could lead to numerous future studies.

Kinyarwanda, like many other languages in contact, has adapted foreign words to meet the needs of its daily life vocabulary and activity. In addition to the lexical need filling, Kinyarwanda borrowed foreign words not only out of need for foreign words but also for prestige. This thesis is based on two hypotheses: Kinyarwanda has borrowed foreign words out of need in various areas; loanwords have been allocated to Kinyarwanda noun class system. This work has discussed and analysed how French and English loanwords have been allocated to key areas of influence and the nominal class system of Kinyarwanda. The data were collected from various sources, including publications, conversation, newspapers, Bible literature, school text books, commercial posters, hoardings. The study has analysed loanwords from French/English deceptive cognates in a bilingual context. This is a challenging task for other researchers who will have to deal with the complexity of deceptive cognate loanwords. / African Languages / M.A. (African Languages)

This study uses a predominantly quantitative approach to explore the effect of different error correction feedback mechanisms on students’ English as a Second Language writing (narrative and descriptive) amongst high school students in Botswana. A longitudinal, quasi-experimental design is used, with a control group that received no correction feedback while the experimental groups received direct, coded and uncoded feedback. Three hypotheses define the study in terms of fluency, correction success and accuracy development over time. No significant increases in fluency were found between the pretests and posttests. Correction success achieved by the three treatment groups when rewriting texts reflected the explicitness of the feedback, with the direct group highest, followed by the coded and uncoded groups. Findings were mixed on the important issue of accuracy development, although they strongly suggest that for spelling, any type of feedback is significantly better than none and that coded feedback is better than direct feedback despite the latter being more explicit. Students from all the treatment groups expressed similarly positive opinions on correction feedback. / Applied Language / M.A. (Spec. in Applied Linguistics)

In higher education institutions, virtual learning systems (VLSs) have been adopted, and are becoming increasingly popular among educators. However, despite this ubiquity of VLS use, there has not been widespread change in pedagogic practice to take advantage of the functionality afforded by VLSs. Knowledge of the actual usage of e-learning systems is limited in terms of what specific feature sets are deemed useful, and how this influences system usage. VLSs have a suite of tools with associated functions/features and properties, as well as non-functional system characteristics. In addition, these systems incorporate pedagogic features to cater for online teaching. Educators in higher education, who are the chief agents of e-learning, are confounded by system-related, pedagogic, organisational, user difference and demographic factors that influence VLS usage. Virtual learning system usage involves system feature usage extent and frequency, total system usage and usage clusters. The aim of this study is to develop a model representing the factors that influence usage of VLSs in higher education. The links between system usage and system-related factors, pedagogic factors, organisational factors, user-difference and demographic factors is researched. This research incorporated a literature study, a pilot study, interviews and surveys. A case study research strategy was combined with a mixed methods research design. The results of the qualitative analysis was triangulated with the findings of the quantitative analysis and compared to the findings of the literature study. The study was conducted at two residential higher education institutions (HEI), namely, University of KwaZulu-Natal and Durban University of Technology. The main contribution of this study is the Virtual Learning System Usage Model (VLSUM) representing the factors that influence VLS usage in residential higher education institutions. The proposed VLSUM is based on the empirical results of this study. VLSUM can be used by managers of educational technology departments and instructional designers to implement interventions to optimize usage. The constructs of VLSUM confirmed existing theories, replicated and synthesised theories from different fields, and extended existing models to produce a new model for understanding the factors that influence VLS usage in higher education. / Computing / D. LITT. et. Phil. (Information Systems)

Objectives: The overall aim of this research was to describe the health-seeking behaviours (HSBs) of people in rural Makwanpur, Nepal, and to analyse the patient, household, community, health-system, knowledge and illness factors, and the psychological, social and cultural processes which explain these behaviours. Background: Much about the health status of populations and individuals can be understood by studying how people utilise their health services and the factors associated with this utilisation. HSB studies act as a starting point for the planning of health programmes and the structuring of health systems. Nepal, with its shortage of health providers and funding, its low service usage and its pluralistic medical landscape provides an interesting setting in which to examine HSBs. Most health policies in this context have been devised without taking into account the perspectives of the system users. Moreover, limited formal research on this topic has been carried out in this context. Methods: This study involved a mixed-methods, explanatory sequential design consisting of two phases – quantitative data collection followed by qualitative data collection. Quantitative data was collected using a cross-sectional household survey carried out in 2,334 households across ten VDCs in Makwanpur district between April 2011 and August 2011. Households were selected using a random sampling method. The survey asked about care-seeking in response to an acute episode of illness in the previous one month. Qualitative data was collected after the quantitative data using semi-structured household interviews (n=90) in three VDCs between November and December 2012. The Qualitative interviews were designed to compliment the quantitative findings and to determine the explicit factors associated with care decisions. Results: Of the 2,334 households surveyed,46% had at least one episode of illness in the month prior. The majority of illnesses were infectious or parasitic diseases (42%). Of those households experiencing illnesses, 69% chose to seek care outside of the home; 22% used traditional healers, 37% used allopathic providers and 12% opted for pharmacies as a first option. Sixteen did nothing to address their illnesses, sighting geography, finances, workload and lack of severity as the reasons. Regression models revealed that a host of different patient, household, community, illness, health facility and knowledge factors were associated with care decisions depending on the decision, but illness factors had the greatest impact overall on whether or not a household sought some care or care outside of the home, while household level factors had the greatest impact on the type of care sought outside of the home and the length of delays before seeking care. Patient gender had an impact on whether or not allopathic care was used at least once. Qualitative results revealed that health facility factors were also equally important in determining households’ conscious decision-making about specific providers. Conclusions: Overall it appears that people in Makwanpur are not underutilising health services as suspected. Households use certified government providers most frequently to address their illnesses, and do not rely too heavily on traditional healers or informal providers exclusively. The results indicate that while illness and household factors are important, in order to improve HSBs and increase allopathic care utilisation, a focus should be on improving health service delivery rather than on changing HSBs. However, in order to decrease delays in care-seeking, a focus should be made to reduce household-level barriers to care as well.

Digital divide research has recently begun to address the functional gaps between Internet-connected technologies, specifically mobile and wired devices. This study uses nationally representative survey data from the Pew Internet and American Life Project to address this area of research and explores how smartphone-dependence compared to multi-modal access impacts Internet use among key demographic groups including race, sex, age, income, and education. This study also explores how demographic characteristics and smartphone use interact to affect reliance on smartphones and perceptions of the utility of mobile devices. Results show that race, sex, age, income, and education, exhibit different rates of smartphone-dependence, and also perform different online activities with their smartphones. Minorities and younger users are more likely to be smartphone-dependent and multi-modal users suggesting that these demographic groups are adopting mobile Internet technologies faster than Whites and older individuals. Minorities also use smartphones for more news and information activities than Whites, which contradicts traditional usage gap predictions.

This essay aims to explore the dialect use in speech and writing in different school contexts among pupils in grade three in Värmland, Sweden and their teacher’s attitude to dialect usage in the classroom. The methods used for the survey are structured observations, collected documents and qualitative semi- structured interviews with both pupils and their teacher. The purpose of using three different methods is to get as complete a survey and credible result as possible as the methods in this case complement each other in an advantageous way. The result shows that a majority of the participating pupils use their dialect in most of the different speech situation with the exception of oral presentations and presentation of today´s date and lunch during morning assemblies. In these situation the dialect is often modified. As for the pupils´ writing in dialect, the result shows, however, that it seldom occurs at all. The pupils generally do not see themselves as dialect speakers while their teacher argues that everyone in the class speaks the local dialect. Furthermore, the results of the teacher interview shows that the teacher is sympathetic to the use of dialect in general and sees no problem with dialect in speech, but does not accept dialect use in school writing. / Min studie syftar till att undersöka värmländska elevers dialektanvändning i såväl tal som i skrift i olika skolsammanhang samt att ta reda på deras lärares inställning till dialektanvändning i klassrummet. Metoderna som legat till grund för undersökningen är strukturerade observationer, insamlade dokument samt kvalitativa semistrukturerade elev- och lärarintervjuer. Avsikten med att använda tre olika metoder är att få en så fullständig undersökning och trovärdigt resultat som möjligt då metoderna i detta fall kompletterar varandra på ett fördelaktigt sätt. Resultatet av undersökningen visar att majoriteten av de medverkande eleverna använder sig av sin värmländska dialekt i de flesta talsituationer, med undantag för muntliga framställningar och presentation av dagens datum och lunch under morgonsamlingar. Vid dessa tillfällen modifieras ofta dialekten. Vad gäller elevernas skrivande på dialekt visar resultatet däremot att det nästan inte förekommer alls. Eleverna själva anser sig överlag inte heller tala värmländska medan deras lärare hävdar att alla i klassen talar dialektalt. Vidare visar resultatet från lärarintervjun att läraren är positivt inställd till dialektanvändning generellt och ser inga problem med dialekt i tal, men accepterar inte dialektal användning i skrift i skolsammanhang.

Aliteracy defines those who can read adequately but who choose not to read for their own interest and pleasure. Adolescent aliteracy is an international issue (OECD, 2000, 2010a). Dissonance between what schools and students consider as ‘engaging reading’ is widening (Ivey & Broadhuss, 2001; Wilheilm & Smith, 2002). Recent evidence of poor literature knowledge amongst teachers and pre-service teachers (Cremin, Mottram, Bearne, & Goodwin, 2008; Nathanson, Pruslow, & Levit, 2008) highlights the need for pragmatic ways to empower adolescent students to address aliteracy for themselves. The aim of this thesis was to explore the conceptual basis for adolescent aliteracy in the 11-13 year old age groups alongside pedagogy to support currently aliterate adolescents. A mixed methods approach used 8 sets of data to explore reading preferences, reading motivations and self-selection behaviours in a mixed and stratifed sample of currently aliterate students over 6 months. The tools were: a reading preference survey, a Title Recognition Test (Cunningham & Stanovich, 1991), the Motivations for Reading Questionnaire, (Wigfield, Guthrie, & McGough, 1996), library observations, student and teacher interviews, library borrowing records and summative reading scores. Data analysis included thematic analysis, multiple regressions, Chi squared, Wilcoxon signed-ranked tests and Spearman’s correlations. Media based titles, magazines and SMS texting were cited as the most popular reading choices. Avid, poor and currently aliterate adolescent readers had significantly distinct motivational and cognitive reading profiles. Exploratory results with a stratified sample of currently aliterate students suggest that taught self-selection strategies significantly increased motivation to read for challenge and for curiosity and decreased motivation to read for reasons of compliance. Amongst currently aliterate adolescents, results suggest significant interaction between reading identity, reading challenge, reading stamina and reading interest.